Metastasis is responsible for 90% of cancer deaths from solid tumors. Despite this reality, the mechanisms of metastasis formation are relatively understudied compared to that of carcinogenesis, deregulation of cellular pathways, or treatment resistance. It stands to reason that focused studies of metastasis provide opportunities to improve survival by identifying aspects for therapeutic development. This proposal will address Provocative Question Group B-6, Given the difficulty of studying metastasis, can we develop new approaches, such as engineered tissue grafts, to investigate the biology of tumor spread? The hypothesis of this proposal is that subclonal genetic evolution within a primary carcinoma selects for prosurvival phenotypes, and the consequence of these phenotypes is metastasis. We propose three iterative Aims to address our hypothesis. In our first aim, we will determine the extent to which subclonal evolution generates intratumoral genetic heterogeneity and metastatic subclones within the primary site. For this aim we will apply deep sequencing of primary and metastatic tissues to determine the genetic basis of metastasis, in association with computational modeling of clonal evolution. In our second aim, we will determine the extent to which subclonal evolution nonrandomly selects for specific genes or pathways during pancreatic cancer progression. This aim will rely upon bioinformatics analyses of sequencing libraries to identify core pathways of significance in tumor progression, coupled with mechanistic studies of metastasis formation based on genes and core pathways identified using a variety of models. In our third aim, we will determine the extent to which non-genetic factors such as EMT, oncogenic stromal signaling or immune infiltration correlates with subclonal evolution in the development of metastatic subclones. This aim will rely on unbiased expression analyses of metastatic subclones in association with high-throughput biomarker screening. The expected overall impact of this proposal will be its potential to cause a paradigm change in our understanding of the genetic basis of tumor progression and metastasis, and to identify the core pathways and functions consistently targeted by subclonal evolution. This is important to know for the ultimate goal of devising therapies that target the metastatic phenotype and that take genetic heterogeneity into account, rather than targeting of specific genes that has shown little advantage thus far.